The History of the Text of Aristophanes Alan H. Sommerstein Preliminary Remarks An interval of twenty-four centuries separates the scripts that Aristophanes wrote for the first performances of his comedies from the texts of those comedies as they appear, for example, in Wilson (2007b). This chapter attempts to trace the chain of transmission that leads from the former to the latter.1 The basic process in this chain is that of copying—by hand, for the first nineteen centuries, and thereafter with mechanical, and very recently with electronic, assistance. Copying, however it is performed, is always liable to error. Those who copy texts are normally aware of this, or at least are supervised by persons who are aware of it and are on the lookout for possible errors in the text they are copying. Often an error will be detected and successfully corrected, thanks either to the copyist‘s or editor‘s own understanding of language, style, and context, or to comparison with another copy that has escaped the error; but there is always also the possibility that an attempted ‗correction‘, far from restoring the text as it was before the error appeared, may actually take it further away from that state,2 or that an ‗error‘ may be detected where the text was in fact sound.3 When all copying is by hand, the net outcome of this process, at most times and in most circumstances, will be a slow increase in the distance between the original and the current state of the text. If this tendency has been reversed in the last half millennium, as on the whole it has, this is due partly to the technology of printing (and later developments that have built upon it), which has both vastly increased the dissemination of texts and reduced the number of separate acts of copying required to effect it, thus putting a virtual halt to the long process of random deterioration; partly (though for most texts, including Aristophanes, only to a rather small extent) to the discovery, mostly in Egypt, in and since the nineteenth century, of fragments of copies far older than any previously known; partly to improvements in communications that have made it possible, as it never was in ancient or mediaeval times, for one editor to have access to virtually all the significant evidence existing in the world that bears on the constitution of the text; and partly to the advancement of our knowledge and understanding of the transmission of texts, the ways in which errors can occur, and the forms they can take. The above remarks apply, with minor variations, to all ancient Greek texts. I now turn to consider the text of Aristophanes in particular. 2 1. The Earliest Days The script of a dramatic performance is inherently unstable. Any text may be altered after its completion as a result of second thoughts by the author; but in a play text, the director,4 the performers, and the audience(s) have also to be considered. The script may be changed during the rehearsal period before the first performance, or afterwards with a view to subsequent performances. The impact of such changes on the eventual dissemination of the script as a reading text may take any of three forms. (i) The change never finds its way at all into the copy or copies of the script from which later reading texts derive. In this case, we will never know about it for sure, except in the unlikely event that a reliable tradition about the actual performance is eventually committed to writing by a later author and either survives in that author‘s text or is quoted therefrom by another surviving writer. (ii) The change is inserted by the author into his working copy. In this case, too, we will usually never know that any change was ever made, but for the converse reason that evidence of the earlier state of the text will normally not survive. Sometimes, however, for special reasons, the alteration will leave visible traces. In Wasps, one of the choral interludes (1265–1291) contains, in the mediaeval manuscripts, a strophe, epirrhema, and antepirrhema, but no lyric antistrophe. The scholia state that something is missing, but that in itself might be merely an inference based on Aristophanes‘ normal practice. However, the first-century metrician Heliodoros5 reports that where the antistrophe should come there were ‗seven lines6 containing dots and marks indicating a corrupt text (ζηιγμὰ ρ καὶ ἀ λόγοςρ), whose sense cannot readily be established‘; he assumed, as he had ‗often said‘ in regard to other similar passages,7 that these lines were already corrupt in ‗the earliest copies‘—too corrupt to be intelligible to later scholars (say, in Hellenistic Alexandria). Since it is unlikely that any merely accidental process (e.g., a damp patch) would so neatly ruin an entire antistrophe while leaving the adjacent epirrhema and antepirrhema untouched, we must suspect that there has been an intentional deletion, and I have suggested (Sommerstein 1983, 233) that Aristophanes himself cut out the song before production, perhaps because someone satirised in it had suddenly died. Either this happened so late that there was no time to write substitute lyrics, or Aristophanes forgot or did not care8 to insert the substitute text in his working script after deleting the original; in either case, all that was left of the antistrophe was a half-erased passage in which, in the words of Heliodoros, it was possible ‗to determine how many lines there were, but not what was in them‘. A late insertion, too, may betray itself by its content 3 and/or by not fitting quite perfectly into its context: an example of the former is the passage in Ecclesiazusae (1154–1162) that refers to the result of the drawing of lots for the order of performance of the competing plays, while the latter is exemplified by the passages in Frogs (71–88, 786–795, 1515–1519) that presuppose, as the rest of the play does not, the death of Sophocles.9 Other changes in Frogs appear to have been made at a later stage, for a repeat performance about a year after the first, and inserted in the working script at the top or bottom of the relevant columns, without obliterating the original text, so that we have been left with a series of doublet passages.10 (iii) If the alterations are extensive, it may become necessary to write out the script again from scratch. This is clearly what was done when Clouds was revised, some years after its first performance, and both the original and the (incompletely) revised script survived to be catalogued and discussed by ancient scholars11 (it was the revised script that made it into late antiquity and is preserved today). More surprisingly, it also seems to have happened when some apparently rather minor revisions12 were made to Wealth for a second production on an unknown occasion; here again the original and revised versions both survived to be studied by at least one later scholar, leading him to the absurd conclusion that the earlier version (the one we have now) was actually the play of the same name that Aristophanes had presented twenty years earlier—in spite of ample internal evidence that it belonged to the period of the Corinthian War.13 In a case like this, the ‗rewriting‘ may have been merely a matter of cutting and pasting (in the literal, not the computer-age, sense) the particular columns of text in which alterations had been made. These various phenomena, especially the apparent deletion in Wasps and the doublets in Frogs, strongly suggest that our texts of the plays derive ultimately from what I have called Aristophanes‘ ‗working scripts‘, which were first written out in fair copy (presumably before rehearsals began) and updated as necessary—and sometimes rewritten—to take account of subsequent changes. We may note, too, that if a text was rewritten, both the earlier and the later versions of it might sometimes go into circulation.14 But that statement raises another question: how did copies of these ‗working scripts‘ get into circulation, for some of them eventually to find their way, a century after the author‘s death, into the catalogue of the Library of Alexandria? 2. The Fourth Century Aristophanes‘ literary heirs will certainly have been his sons, Philippos and Araros15—both of whom, like so many sons and other relatives of great Athenian dramatists, followed him 4 into the same profession. Philippos, who was given his paternal grandfather‘s name, was presumably the elder, but their father evidently thought Araros the more talented16 and entrusted to him the production of the last two plays he wrote.17 We are told, however, that Araros did not produce any of his own plays at Athens until some time in the 101st Olympiad (i.e., between 375 and 372, inclusive);18 I suspect that for a considerable part of the intervening period, he may have been making a good living by producing his father‘s comedies in other parts of the Greek world, especially in the west.19 Philippos meanwhile remained at home, looking after the family property and from time to time composing comedies himself (he won first prize at the Lenaia at a date not later, and probably a little earlier, than 378).20 Araros will of course have taken copies of his father‘s scripts with him on his travels, but it would make sense for Philippos to retain the originals at Athens and make them available for copying, whether as a free service to friends or as a paid one to booksellers.